Replicate was founded in 2019 by Ben Firshman (creator of Docker Compose) and Andreas Jansson in San Francisco. The platform makes it simple to run machine learning models through an API — developers push a model to Replicate, and it handles serving, scaling, and billing automatically.
Replicate raised over $60 million, including a $40 million Series B led by Andreessen Horowitz in 2024 at a $350 million valuation. The company’s model marketplace has become a popular way for researchers to share working models and for developers to access them.
The platform hosts thousands of open-source models, from Stable Diffusion and LLaMA to specialized models for audio, video, and 3D generation. Developers can run any model with a single API call, paying only for the compute time used. Replicate also supports custom model deployment using Cog, their open-source tool for packaging ML models.
Replicate has become the go-to platform for rapid prototyping with open-source AI models. Its pay-per-second pricing and instant API access let developers test ideas without upfront GPU costs or infrastructure setup.
Over 100,000 developers use Replicate, and the platform runs millions of predictions daily. Notable community models on the platform have been run billions of times collectively. The company employs around 50 people and continues to expand model support and infrastructure capacity.
